Tuesday, December 28, 2010

Applied Sraffianism

"These allusions give incidentally some indication of the disproportionate length of time over which so short a work has been in preparation... As was only natural during such a time period, others have from time to time independently taken up points of view which are similar to one or the other adopted in this paper and have developed them further or in different directions from those pursued here." -- Piero Sraffa

Despite my appreciation of Bliss' 1975 book, I think the following view dubious, uninformed, and authoritarian:
"A striking feature of the school to which Piero Garegnani belongs is its seeming lack of interest in the real world... Our world is changing rapidly and in ways that demand economic analysis of what is happening... Over the last 30 years so-called neoclassical economics has been extraordinarily productive... What has been the contribution of the post-Sraffa school in the same period? Nothing at all as far as I can see. This has been an exceptionally sterile approach. Where are the new ideas? Where are the illuminating insights into what is happening today?" -- Christopher Bliss (2009)
I have no problem with a researcher deciding to center their investigations into criticism and the history of economic thought. I think that when Bliss calls neoclassical economics "extraordinarily productive", he includes research that fails to test neoclassical economics and whose relationship to neoclassical economics can be doubted. Sraffian economics can easily exceed this standard.

I take the "others" Sraffa refers to above to be principally Wassily Leontief and John Von Neumann. So, for example, work with Leontief's Input-Output (I/O) analysis is applied Sraffian analysis. Interestingly enough, countries maintain their national accounts in a form supporting I/O analysis. (For the United States, see the benchmark input-output accounts available from the Bureau of Economic Analysis (BEA). For many developed countries, see the Structural Analysis (STAN) Database for data on Industry and Services available from the Organisation for Economic Co-Operation and Development (OECD).) For me, challenges in working with this data arise from statistical discrepancies, rectangular matrices that I expect to be square, components in value added that are neither wages nor profits, import and export flows, etc. But other economists, including some Sraffians have addressed these challenges in their own work.

I have listed selected applied Sraffa work on two topics.. Tony Aspromourgos (2004) lists Sraffian research applied to a larger range of issues.

References
  • Tony Aspromourgos (2004). "Sraffian Research Programmes and Unorthodox Economics", Review of Political Economy, V. 16, n. 2: pp. 179-206.
  • Christopher Bliss (2009). "Comment on 'Capital in Neoclassical Theory: Some Notes' by Professor Piero Garegnani".
  • Thijs ten Raa (2006). The Economics of Input-Output Analysis, Cambridge University Press.
  • John Von Neumann (1945). "A Model of General Economic Equilibrium", Review of Economic Studies, V. 13, N. 1: pp. 1-9.

11 comments:

Anonymous said...

A narrow post-Sraffian approach can hold its own against neoclassical economics. But the real standard of value is to what extent a narrow post-Sraffian approach can hold its own against economic reality. And here, I think it's fairly clear, that a narrow post-Sraffian approach is inherently static and deterministic, and not significantly different from neoclassical models.

Isn't the best critique the formulation of an alternative theoretical framework that better explains economic reality? Do you really think linear algebra is the right tool to understand capitalism?

BruceMcF said...

A key distinction between neoclassical and Sraffian approaches with respect to understanding economic reality is that the Sraffian system is analytically open, and so is compatible with empirically grounded models of human decision making, unlike neoclassical economics which is built on a falsified model of human decision making.

Since capitalism cannot be encompassed in mathematical modeling alone, mathematical approaches that are analytically open offer the prospect of being a useful tool for cause and effect explanations of capitalism, where neoclassical economics can aspire to no better than emulation and simulation of capitalism.

Anonymous said...

Robert, check out Pivetti´s and Panico´s work on the influence of interest rate on the rate of profit and prices. It´s full of policy analysis (for instance, the subprime crisis)

Gualra

Anonymous said...

I think the fact that Sraffa's system is "analytically open" is not of great interest or import. Yes, it's an under-determined simultaneous equation system with two degrees of freedom. You can close it by specifying an exogenous distribution of income and a numeraire. But this property pales into significance compared to the complete absence of the analysis of out-of-equilibrium dynamic adjustment of a capitalist economy. The Sraffian system has all the faults of the neoclassical system in this respect -- it is static, or, at best, agnostic about economic change.

I think it's incorrect to claim that because Sraffa's system is under-determined it is therefore a "useful tool for cause and effect explanations". You might get comparative static "cause and effect" stories out of it, but, again, without explicit out-of-equilibrium dynamics such comparative statics exercises are without foundation.

BruceMcF said...

You contradict yourself: first you say that the Sraffian system has all the faults of the neoclassical system, and then you say that it is at best agnostic about change.

I agree with the second: the fact that the Sraffian system is a genuinely static system means that it can be agnostic about change. It is not a model that demands that it be set in historical time, but it is a model that is compatible with being set in historical time. The neoclassical system, by contrast, is incompatible with being set in historical time.

Anonymous said...

Both neoclassical models and (narrow) Sraffian models are systems of simultaneous equations *completely lacking* any specification of out-of-equilibrium dynamic adjustment. On this vital axis they possess a strong family resemblance. And -- I hope you would agree -- this simply isn't good enough from the point of view of understanding economic systems.

You claim that -- because Sraffa's formulation is under-determined, i.e. leaves the distribution of income undetermined -- it is "compatible with being set in historical time" in contrast to neoclassical models. But I do not see how these properties are connected. Any static model is in principle compatible with being set in historical time because the static model may be interpreted as the linearisation of the full dynamic system at an equilibrium point.

Perhaps the contrast you are alluding to is that neoclassical general equilibrium assumes tatonnement? i.e. trade cannot take place out-of-equilibrium? Agreed, Sraffianism is not as off-the-wall deranged as that.

Anonymous said...

"Sraffian" systems deals with long-period positions of economic variables; those are influenced by strong and persistent forces. In fact there is a close resemblace with traditional neoclassical theory: Marshall, Wicksell and Walras share this idea of persistence. The main difference is the content of the theory (¿which are those forces affecting long-term outcomes?)

Gualra

Robert Vienneau said...

First anonymous, please create a pseudonym in future comments.

I think we all know that there's an increasing amount of literature arguing whether or not Sraffa's approach is consistent with models set in historical time. I think my finding the Sraffa critique of neoclassical economics convincing inclines me to want to see a constructive Sraffian approach, aside from an interpretation of the history of economics. I would like to accept that Sraffa's equations make sense without a formal specification of a dynamical system in which they are limit points. Ajit Sinha's position is interesting. I also wonder if one could talk about a discovery process, akin to Hayek's. I'd like to avoid being mystical, though.

Part of the challenge is specifying the data for Sraffa's model such that the long period does not have an independent existence, over and above a succession of short runs. According to Sraffa, his data do not require any assumption on returns to scale. That implies, I guess, the level of gross outputs must be given. How can this be consistent with levels being those for a pre-existing long run position? Are capacities being operated at actual levels for these data, planned levels, or what?

Gualra, thanks for the recommendation. I did think about citing Pivetti and Panico's separate contributions to a Garegnani festschrift. I have never had access to Pivetti's 1991 or Panico's 1988 books. So I don't what a full statement of their positions looks like. I gather, however, they rely on the open character of Sraffa's approach that Bruce mentions.

I don't mind if you continue to argue about Davidson versus Garegnani or whatever here. But please keep in mind that a comment section can probably not do justice to any positions in the spectrum. Did the Trieste summer school discussions end up generating lots of bad feeling?

Anonymous said...

Robert, check out Franklin Serrano´s work. It´s mainly "Sraffian"

http://www.ie.ufrj.br/ecopol/franklin.php

cheers,
Gualra

Revista Circus said...

Robert,

From Serrano´s work, was opened a link between Kalecki and Keynes to long period positions of classical theory, and therefore applied sraffian dynamics models with a separation of prices and quantities.

Here the sraffian supermultiplier of Franklin Serrano, (in Bortis there is another version of supermultiplier,

https://sites.google.com/site/serranofranklin2/home/Serrano-1995_Sraffiansupermultiplier.pdf?attredirects=0&d=1

more recently Olivera, M. "Effective Demand, Economic Growth and External Constraints: Rethinking Regional Integration in Latin America"
US: http://www.amazon.com/Effective-Demand-Economic-External-Constraints/dp/383838296X/ref=sr_1_6?ie=UTF8&s=books&qid=1282579009&sr=8-6
):

Gravitation of market-prices over normal-prices, represents the dynamics of prices within the core.
another recent paper of this topic is:
Bellino-Serrano: "Gravitation analysis:beyond cross-dual models and back to Adam Smith"

ale.

Anonymous said...

The work of Bellino etc. analyzing the dynamics of gravitation is precisely the kind of work that is needed to get beyond the narrow post-Sraffianism that remains stuck on static, linear i/o models of the economy.

Some Sraffians claim to uphold the Classical tradition in Political Economy. But to really walk in the shoes of Smith, Marx etc. requires a truly dynamic analysis of capitalism. 90% of the post-Sraffian work is static. I think this needs to change.